{"id":"8ee223ed-8ba3-4206-8dae-bad7c18db24c","arxiv_id":"2501.13149","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Standard infall models are unreliable for individual galaxy radial velocities, but their velocity dispersions provide useful upper and lower bounds.","lead":"Astronomers measure only the line-of-sight velocity of most galaxies, and use 'infall models' to estimate how galaxies rush toward each other. This paper generalizes those models, tests them on more than 5,000 simulated galaxies, and finds they fail to predict individual velocities but are still useful for measuring velocity dispersions.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Abstract's M81 upper/lower-bound labels are reversed relative to the paper's own §4 finding, and the quoted debiased dispersions (180, 142) do not follow from the stated m_σ,b_σ fits; the headline M81 prediction is not reproducible.","rationale":"The paper's analytical generalization of the infall models and its simulation-based demonstration that individual infall velocities are not reliable are genuine contributions. The reader's conditional verdict is appropriate because the core methodological finding may survive a correction of the M81 application. However, the single most load-bearing problem is not the representativeness of the simulated M81-like halos per se, but the internal inconsistency in how the M81 prediction is produced and labeled: the abstract reverses the lower/upper-bound roles established in §4, and the reported bias-corrected values cannot be reproduced from the quoted m_σ,b_σ fits by any standard inversion. This directly affects the headline M81 numbers, which are the paper's central advertised result. The concrete test—recomputing the inversion and checking bound ordering—would settle whether the discrepancy is real or due to an undocumented calibration step. Because the analytical core and the general simulation finding remain plausible, the appropriate verdict stays conditional: the M81 section needs a corrected, reproducible calibration before its specific numbers are used.","tokens_in":21425,"tokens_out":11352,"duration_ms":121208,"concrete_test":"Recompute the M81 bias-corrected dispersions from the observed values in Eqs. (60)–(62) and the §5.1 fits by solving σ_obs = m_σ σ_true + b_σ for each model, propagating the quoted 1σ fit bounds. Then compare the resulting intervals with the abstract's (180±42), (142±64), and (99±36), and check that the bound ordering matches §4's statement that minor underestimates and major overestimates. If the recomputed values differ by more than the quoted uncertainties or the ordering is reversed, the M81 prediction is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that infall-model velocity dispersions provide robust bounds is applied to M81 in a way that is internally inconsistent. In §4 and §5.1, the linear fits show that the minor infall model underestimates the true dispersion (m≈0.5, b≈22–26 km/s) while the major infall model overestimates it (m≈3.5–4, b≈208–255 km/s). The raw M81 dispersions are σ_min=96, σ_maj=564, σ_Δv=102 km/s (Eqs. 60–62). Under the natural inversion σ_true = (σ_obs−b)/m, the §5.1 fits give roughly 140, 77, and 63 km/s, not the abstract's (180±42), (142±64), and (99±36). Moreover, the abstract labels the minor-infall estimate an 'upper bound' and the major-infall estimate a 'lower bound,' the exact reverse of the §4 statement that minor systematically underestimates and major systematically overestimates the true dispersion. Unless an unreported regression or calibration procedure produced the quoted values, the M81 prediction is not derivable from the presented analysis. This does not disprove the general kinematics derivation, but it undermines the paper's headline application and the advertised 'robust upper and lower bounds' claim for M81.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper generalizes the minor and major infall models of Karachentsev & Kashibadze (2006) to arbitrary velocity directions, deriving exact geometric expressions for the radial and tangential components of the relative velocity between two galaxies in terms of observable line-of-sight components, distances, and angular separation. The authors test the models on Illustris-3 simulated halos, finding that the infall-model velocities themselves reproduce the true radial velocity only for a minority (~34%) of subhalos, whereas the velocity dispersions computed from the infall-model velocities bracket the true dispersion for most halos. They then apply the models to the M81 group, quoting debiased radial-velocity dispersions of (180±42), (142±64), and (99±36) km/s from the minor, major, and line-of-sight-difference estimators.","tokens_in":21734,"tokens_out":6615,"duration_ms":59371,"significance":"The purely kinematical derivation in Sect. 2 is a valuable contribution: it shows that the two classical infall models correspond to different projections of the same relative-velocity vector and that they are cosmology-independent. The simulation analysis provides a concrete, quantitative warning that perpendicular and tangential velocity components cannot be neglected even at small angular separations, and the demonstration that the infall-model dispersions bracket the true dispersion is a practically useful result for group/cluster studies. The paper also makes a falsifiable prediction for the M81 group. However, the M81 prediction is currently presented with internally inconsistent bound labels and non-reproducible numbers, so a revision is needed before the headline claim can be accepted.","major_comments":[{"comment":"The abstract labels the minor-infall estimate as an 'upper bound' and the major-infall estimate as a 'lower bound' on the M81 radial-velocity dispersion, but Sect. 4 and Fig. 7 establish the opposite bias: the minor infall model systematically underestimates the true dispersion and the major infall model systematically overestimates it. If the raw infall-model dispersions are to serve as bounds, the minor model provides the lower bound and the major model the upper bound. This reversal must be corrected in the abstract and in the conclusions.","section":"Abstract and Sect. 5.2"},{"comment":"The quoted debiased dispersions (180±42, 142±64, 99±36 km/s) are not derivable from the stated linear fits in Fig. 10 (centre). Using the natural inversion sigma_true = (sigma_obs - b)/m with the given fits (minor: m=0.53, b=22.21 km/s; major: m=4.01, b=254.60 km/s; Eq. (46): m=1.03, b=37.31 km/s) yields approximate values of 139, 77, and 63 km/s, not 180, 142, and 99. The quoted values are close to sigma_obs/m, which ignores the intercept b; if that is the intended bias correction, it should be stated explicitly with proper error propagation from both m and b. As written, the M81 prediction is not reproducible from the presented fits.","section":"Sect. 5.2, Eqs. (63)-(65)"},{"comment":"The Bayesian re-analysis in the same section gives different central values (150, 88, 100 km/s for minor, major, and Delta-v models) from the abstract's (180, 142, 99). The paper does not explain which set of numbers is the final prediction, nor does it reconcile the two. This ambiguity affects the headline claim and needs to be resolved.","section":"Sect. 5.2, Bayesian estimates"}],"minor_comments":[{"comment":"The description of the linear fit '|v_inf| = m |v_r| ... with its 1-sigma confidence bounds b_v' is ambiguous: are the b_v values (81.06 km/s etc.) intercepts or scatter bounds? If intercepts, include them in the equation; if bounds, report them differently. This ambiguity propagates to the calibration fits in Sect. 5.","section":"Sect. 4, linear fit description"},{"comment":"The text reports a major-infall fit (m_v=1.01, b_v=5578.32 km/s) but states the fit is not shown because its confidence bound covers the entire plot; this should be clarified, since the reported numbers are otherwise untestable.","section":"Fig. 5 (left)"},{"comment":"The M81-like halo selection does not explicitly select for merging groups with strong interactions like M81/M82/NGC 3077; the authors note that the simulated structures may not fully represent the M81 group, but a more quantitative discussion of this extrapolation (e.g., using the 100 kpc offset criterion as a proxy for the merging state) would strengthen the application.","section":"Sect. 5.1"},{"comment":"The phrase 'for more than 90% of all, more than 5000 infalling subhalos' is grammatically awkward and should be rephrased for clarity.","section":"Abstract"},{"comment":"The statement 'Amplitudes |.| can be negative' is confusing because the notation |.| is normally reserved for non-negative magnitudes; the paper already uses ||.||_2 for the latter, so consider using a different symbol for the signed amplitude.","section":"Sect. 2.1"}],"recommendation":"major_revision","confidential_remarks":"The core kinematical derivation and the simulation-based warning about the infall velocities are solid and worth publishing. However, the M81 application in its present form contains a sign/label inconsistency and an unreproducible numerical estimate; the authors need to revise this section carefully and state their calibration procedure unambiguously. As a referee I would not recommend rejection, since the issues are local and fixable, but they must be addressed before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the general infall-model derivation in Sect. 2 is a genuine step forward, and the simulation result on velocity dispersions is robust and useful. But the M81 application is not in a publishable state: the abstract's upper/lower labels are reversed relative to the body's own finding, and the quoted debiased dispersions (180, 142, 99) do not follow from the stated linear fits. That needs to be fixed before I'd trust the numbers.\n\nWhat's new: the paper derives the full pairwise radial velocity without dropping perpendicular and tangential components, gives the conditions (Eq. 32) under which minor/major over- or underestimate, and shows the two models coincide only for theta=0 or pi (plus a trivial identity at small angles). The Illustris-3 analysis is clean: for 5,036 subhalos around 344 halos, only ~34% have true radial velocity bracketed by the infall models, yet the model dispersions do bracket the true dispersion in most halos. That distinction -- velocities unreliable, dispersions usable -- is the paper's real contribution and it is well supported.\n\nSoft spots, in order of importance. (1) M81 calibration internally inconsistent. The abstract calls minor an upper bound and major a lower bound, but Sect. 4 and the fits show the opposite: minor underestimates, major overestimates. Using the stated fits m_sigma, b_sigma to invert the raw dispersions (96, 564, 102) gives roughly 139, 77, 63 km/s -- not 180, 142, 99. The Bayesian procedure in Sect. 5.2 gives 150, 88, 100, still not the abstract values. So the headline numbers are not reproducible from the text as written. (2) The 3000 km/s cutoff for major-infall velocities is ad hoc; dropping 2% of subhalos at very high velocity could bias the dispersion fits. The effect is said to be marginal, but it should be justified or varied. (3) The M81-like halo selection is reasonable but the posterior depends on simulation representativeness; the authors are honest about the sparse statistics.\n\nBottom line: the kinematic derivation and the dispersion-bracketing result deserve a serious referee. The M81 section needs a rewrite -- fix the bound labels, show how the quoted numbers were obtained, and probably report the Bayesian or inverted values consistently. I'd send it to review, and expect major but fixable revision.","headline":"Solid kinematic generalization and a useful dispersion-bracketing result, but the M81 application is internally inconsistent and the quoted numbers don't follow from the stated fits.","tokens_in":22243,"tokens_out":4042,"would_cite":false,"duration_ms":37635,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Infall-model velocities misestimate individual radial velocities for most galaxies, yet the dispersions they yield systematically bracket the true velocity dispersion, predicting the M81 group's at 99–180 km/s.","keywords":["radial infall velocity","minor infall model","major infall model","velocity dispersion","line-of-sight velocity","Illustris simulation","galaxy groups","M81 group"],"falsifier":"Measure the full three-dimensional velocities of M81-group members, for instance proper motions of its dwarf galaxies from high-precision astrometry, compute the true radial velocity dispersion of the group, and check whether it falls inside the predicted 142–180 km/s bracket and near the $(99 \\pm 36)$ km/s line-of-sight-difference estimate. A cheaper falsification is to rerun the dispersion calibration on a higher-resolution simulation, such as Illustris-1 or TNG, with the same M81-like selection: if the fitted slopes and intercepts move by more than their stated $1\\sigma$ confidence bounds, the M81 prediction loses its calibration anchor.","tokens_in":21178,"feed_emoji":"🔭","tokens_out":18008,"duration_ms":149969,"temperature":0.7,"pith_summary":"The paper asks whether the two standard infall models — the minor and major models used to turn measured line-of-sight velocities into the radial velocity between a galaxy and a group — deliver what observers need. Working from the exact kinematics of a binary motion, the authors show that the two recipes are just different projections of one cosmology-independent velocity identity, and that both are accurate only under a fine-tuned ratio condition involving the unobservable velocity components perpendicular to the line of sight. Tested on more than 5000 subhalos around 344 halos in the Illustris-3 simulation, the individual model velocities fail: the true radial velocity lies between the two model estimates for only about 34% of subhalos, even at angular separations below $10^\\circ$. The positive finding is that the scatter statistics survive — the velocity dispersions computed from the model velocities bracket the true dispersion for most halos, and for distant systems the plain difference of line-of-sight velocities reproduces the true dispersion almost exactly — which is why the authors can still predict the M81-group radial velocity dispersion and argue that dispersion-based mass estimates remain trustworthy.","feed_headline":"Infall models fail on single velocities but bound the dispersion","feed_subtitle":"Most radial-infall estimates miss; their scatter brackets the true dispersion and predicts M81's spread.","key_machinery":"The load-bearing object is the exact relative-velocity decomposition of two galaxies into radial and tangential parts: projecting $\\mathbf{v}_2 - \\mathbf{v}_1$ onto the connection line $\\hat{\\mathbf{r}}_{21}$ gives Eq. (22), which contains the observable line-of-sight components, the angular separation $\\theta$, the distances $r_1, r_2$, and the unobservable perpendicular speeds $|v_{\\perp 1}|$ and $|v_{\\perp 2}|$. Two derived conditions carry the argument. First, the fine-tuning condition $|v_{\\perp 1}|/|v_{\\perp 2}| = |r_1|/|r_2|$ is the only generic situation in which the perpendicular terms vanish, and it fixes whether the minor and major models over- or underestimate the true radial velocity. Second, the small-angle identity $|v_r| \\approx |v_{l2}| - |v_{l1}|$ (Eq. 46) makes all models coincide for small $\\theta$, and in the simulation it reproduces the true velocity dispersion almost exactly for distant halos, converting a kinematic approximation into a practical dispersion estimator.","core_discovery":"The paper's central claim is that the minor and major infall models are not rival physical pictures of galaxy motion but two projections of a single, cosmology-independent kinematics identity for the relative velocity $\\mathbf{v}_2 - \\mathbf{v}_1$ of two galaxies: the minor model projects the line-of-sight velocity components onto the connection line $\\hat{\\mathbf{r}}_{21}$, while the major model projects onto one galaxy's line of sight and divides by a distance factor. Because the exact radial-velocity expression contains the unobservable perpendicular components $|v_{\\perp 1}|$ and $|v_{\\perp 2}|$, each model is accurate only under the fine-tuned ratio condition $|v_{\\perp 1}|/|v_{\\perp 2}| = |r_1|/|r_2|$, which real subhalos essentially never satisfy: in Illustris-3, more than 90% of the 5036 subhalos have non-negligible perpendicular or tangential velocity, and only 34% have their true radial velocity bracketed by the two model estimates. The constructive part of the claim is that the velocity dispersions built from the model velocities are systematically related to the true dispersion — the minor model underestimates it, the major model overestimates it, and the small-angle difference of line-of-sight velocities, $|v_{l2}| - |v_{l1}|$, matches it almost one-to-one for halos beyond 18 Mpc — so that linear calibrations yield reliable brackets on the true dispersion. Applied to M81-group observations, the calibrated relations give a radial velocity dispersion of $(180 \\pm 42)$ km/s from the minor model, $(142 \\pm 64)$ km/s from the major model, and $(99 \\pm 36)$ km/s from the line-of-sight difference, with the last identified as the most likely value because the M81 group satisfies the small-angle condition.","pith_inferences":["A decisive external test is available: for any nearby group whose member velocities can be measured in three dimensions, whether by proper-motion surveys or high-resolution zoom simulations, one can check that the dispersion-bracket property holds outside Illustris-3 and that the M81 prediction of roughly 99–180 km/s contains the measured dispersion.","Because the fitted dispersion calibrations come from one simulation suite and one halo-selection recipe, their transfer to real groups is a testable hypothesis; rerunning the calibration with a higher-resolution simulation or a different halo finder would show whether the slopes and intercepts are stable or selection-dependent.","The paper's kinematics-only treatment suggests the bracket property is cosmology-independent; a natural check is whether the same dispersion brackets appear in simulations with modified gravity or different expansion histories, where the Hubble-flow term the models assume would differ.","If the line-of-sight-difference dispersion remains unbiased for all small-angle systems, virial masses of high-redshift clusters derived from single-spectroscopy surveys would rest on firmer footing; a comparison against X-ray or Sunyaev-Zeldovich mass estimates could test this directly."],"forward_implications":["Minor- and major-model velocity dispersions bracket the true velocity dispersion for most simulated halos, so dispersion-based quantities such as virial masses can be bounded even where individual infall velocities are untrustworthy.","For structures far enough from the observer that $\\theta < 10^\\circ$, the velocity dispersion inferred from the difference of line-of-sight velocities alone coincides with the true dispersion, justifying this shortcut for high-redshift groups and clusters.","The M81-group's radial velocity dispersion is predicted at $(99 \\pm 36)$ km/s from the line-of-sight difference, with the two infall models providing a conservative bracket of roughly 142–180 km/s.","Selecting galaxies at large distances or along the line of sight in front of and behind a cluster centre does not suppress the perpendicular and tangential velocity components, so individual infall-model velocities cannot be used for precise radial-velocity work.","The major infall model's motivation from a vanishing total angular momentum fails for individual halos; the model retains its value as a dispersion-bounding estimator rather than an exact velocity predictor."],"supporting_citations":[{"why":"Establishes the minor and major infall models that the paper generalizes, tests against simulations, and reframes as two projections of one kinematics identity.","marker":"Karachentsev & Kashibadze (2006)"},{"why":"Presents the Illustris simulation whose z=0 snapshot (Illustris-3) supplies the 344 halos and 5036 subhalos used to test the infall models.","marker":"Vogelsberger et al. (2014)"},{"why":"Provides the Illustris public data release and halo/subhalo catalogues from which the relaxed, isolated halos and subhalos are selected.","marker":"Nelson et al. (2015)"},{"why":"Carries the prior claim that true radial velocities lie between minor and major infall estimates; the paper derives the condition under which this fails and finds it holds for only 34% of subhalos.","marker":"Kim et al. (2020)"},{"why":"Supplies the homogeneous M81-group member dataset (positions, distances, heliocentric velocities) used for the observed dispersion prediction.","marker":"Müller et al. (2024)"},{"why":"Provides the M81 and M82 velocities missing from the member catalogue and the Local-Group-centroid frame corrections discussed in the analysis.","marker":"Karachentsev et al. (2013)"},{"why":"Gives the zero-velocity-radius mass relation used to select isolated halos and to define the Hubble-flow subhalo spheres.","marker":"Peirani & de Freitas Pacheco (2006)"},{"why":"Exemplifies the line-of-sight selection strategy (theta approximately 0) for the Virgo cluster that the paper shows does not suppress tangential velocities.","marker":"Karachentsev & Nasonova (2010)"}],"fun_headline_variants":["Infall models: poor for single pairs, robust for dispersion bounds","Infall model projections bound true radial velocity dispersion","Dispersion from infall models brackets true value; single estimates off","M81 group: infall models predict dispersion between 142 and 180 km/s"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the 281 Illustris-3 halos selected to mimic the M81 group — matched in mass, distance, subhalo count, and observer geometry — are representative enough of the real, actively merging M81 group that the simulated linear calibration between infall-model dispersion and true dispersion carries over to the observed M81 velocities.","fun_headline_variants_meta":{"raw":{"variants":["Infall models: poor for single pairs, robust for dispersion bounds","Infall model projections bound true radial velocity dispersion","Dispersion from infall models brackets true value; single estimates off","M81 group: infall models predict dispersion between 142 and 180 km/s"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000795,"raw_usage":{"total_tokens":3680,"prompt_tokens":1306,"completion_tokens":2374,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":922,"completion_tokens_details":{"reasoning_tokens":2298}},"tokens_in":922,"tokens_out":2374,"duration_ms":19702,"temperature":1.0,"reasoning_tokens":2298,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T16:25:15.915506+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the full three-dimensional velocities of M81-group members, for instance proper motions of its dwarf galaxies from high-precision astrometry, compute the true radial velocity dispersion of the group, and check whether it falls inside the predicted 142–180 km/s bracket and near the $(99 \\pm 36)$ km/s line-of-sight-difference estimate. A cheaper falsification is to rerun the dispersion calibration on a higher-resolution simulation, such as Illustris-1 or TNG, with the same M81-like selection: if the fitted slopes and intercepts move by more than their stated $1\\sigma$ confidence bounds, the M81 prediction loses its calibration anchor.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the minor and major infall models that the paper generalizes, tests against simulations, and reframes as two projections of one kinematics identity."},{"cited_title":"2015, Astronomy and Computing, 13, 12 Peñarrubia, J., Ma, Y .-Z., Walker, M","cited_arxiv_id":null,"evidence_quote":"Provides the Illustris public data release and halo/subhalo catalogues from which the relaxed, isolated halos and subhalos are selected."},{"cited_title":"J., Kang, J., Lee, M","cited_arxiv_id":null,"evidence_quote":"Carries the prior claim that true radial velocities lie between minor and major infall estimates; the paper derives the condition under which this fails and finds it holds for only 34% of subhalos."},{"cited_title":"& de Freitas Pacheco, J","cited_arxiv_id":null,"evidence_quote":"Gives the zero-velocity-radius mass relation used to select isolated halos and to define the Hubble-flow subhalo spheres."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Exemplifies the line-of-sight selection strategy (theta approximately 0) for the Virgo cluster that the paper shows does not suppress tangential velocities."}],"review_version":1}